/*=========================================================================
 *
 *  Copyright Insight Software Consortium
 *
 *  Licensed under the Apache License, Version 2.0 (the "License");
 *  you may not use this file except in compliance with the License.
 *  You may obtain a copy of the License at
 *
 *         http://www.apache.org/licenses/LICENSE-2.0.txt
 *
 *  Unless required by applicable law or agreed to in writing, software
 *  distributed under the License is distributed on an "AS IS" BASIS,
 *  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 *  See the License for the specific language governing permissions and
 *  limitations under the License.
 *
 *=========================================================================*/
#ifndef itkBayesianClassifierImageFilter_h
#define itkBayesianClassifierImageFilter_h

#include "itkVectorImage.h"
#include "itkImageToImageFilter.h"
#include "itkMaximumDecisionRule.h"
#include "itkImageRegionIterator.h"

namespace itk
{
/** \class BayesianClassifierImageFilter
 *
 * \brief Performs Bayesian Classification on an image.
 *
 * \par Inputs and Outputs
 * The input to this filter is an itk::VectorImage that represents pixel
 * memberships to 'n' classes. This image is conveniently generated by the
 * BayesianClassifierInitializationImageFilter. You may use that filter to
 * generate the membership images or specify your own.
 *
 * \par
 * The output of the filter is a label map (an image of unsigned char's is the
 * default.) with pixel values indicating the classes they correspond to. Pixels
 * with intensity 0 belong to the 0th class, 1 belong to the 1st class etc....
 * The classification is done by applying a Maximum decision rule to the posterior
 * image.
 *
 * \par Parameters
 * The filter optionally allows you to specify a prior image as well. The prior
 * image, if specified must be a VectorImage with as many components as the
 * number of classes. The posterior image is then generated by multiplying the
 * prior image with the membership image. If the prior image is not specified,
 * the posterior image is the same as the membership image. Another way to
 * look at it is that the priors default to having a uniform distribution over
 * the number of classes.
 * Posterior membership of a pixel = Prior * Membership
 *
 * \par
 * The filter optionally accepts a smoothing filter and number of iterations
 * associated with the smoothing filter.
 * The philosophy is that the filter allows you to iteratively
 * smooth the posteriors prior to applying the decision rule. It is hoped
 * that this would yield a better classification. The user will need to plug
 * in his own smoothing filter with all the parameters set.
 *
 * \par Template parameters
 * InputVectorImage, datatype of the output labelmap, precision of the posterior
 * image, precision of the prior image.
 *
 * \author John Melonakos, Georgia Tech
 *
  * \note
 * This work is part of the National Alliance for Medical Image Computing
 * (NAMIC), funded by the National Institutes of Health through the NIH Roadmap
 * for Medical Research, Grant U54 EB005149.
 *
 * \sa VectorImage
 * \sa BayesianClassifierInitializationImageFilter
 * \ingroup ClassificationFilters
 * \ingroup ITKClassifiers
 */
template< typename TInputVectorImage, typename TLabelsType = unsigned char,
          typename TPosteriorsPrecisionType = double, typename TPriorsPrecisionType = double >
class ITK_TEMPLATE_EXPORT BayesianClassifierImageFilter:
  public ImageToImageFilter<
    TInputVectorImage, Image< TLabelsType,
                               TInputVectorImage ::ImageDimension > >
{
public:
  /** Standard class typedefs. */
  typedef BayesianClassifierImageFilter Self;
  typedef ImageToImageFilter<
    TInputVectorImage,
    Image< TLabelsType, TInputVectorImage::ImageDimension > > Superclass;

  typedef SmartPointer< Self >       Pointer;
  typedef SmartPointer< const Self > ConstPointer;

  /** Method for creation through the object factory. */
  itkNewMacro(Self);

  /** Run-time type information (and related methods). */
  itkTypeMacro(BayesianClassifierImageFilter, ImageToImageFilter);

  /** Input and Output image types. */
  typedef typename Superclass::InputImageType InputImageType;

  /** Dimension of the input image. */
  itkStaticConstMacro(Dimension, unsigned int,
                       InputImageType ::ImageDimension);

  typedef Image< TLabelsType,
                 itkGetStaticConstMacro(Dimension) >        OutputImageType;
  typedef typename InputImageType::ConstPointer InputImagePointer;
  typedef typename OutputImageType::Pointer     OutputImagePointer;
  typedef typename InputImageType::RegionType   ImageRegionType;

  /** Input and Output image iterators. */
  typedef ImageRegionConstIterator< InputImageType > InputImageIteratorType;
  typedef ImageRegionIterator< OutputImageType >     OutputImageIteratorType;

  /** Pixel types. */
  typedef typename InputImageType::PixelType  InputPixelType;
  typedef typename OutputImageType::PixelType OutputPixelType;

  /** Image Type and Pixel type for the images representing the Prior
   * probability of a pixel belonging to  a particular class. This image has
   * arrays as pixels, the number of elements in the array is the same as the
   * number of classes to be used. */
  typedef VectorImage< TPriorsPrecisionType,
                       itkGetStaticConstMacro(Dimension) >                  PriorsImageType;
  typedef typename PriorsImageType::PixelType         PriorsPixelType;
  typedef typename PriorsImageType::Pointer           PriorsImagePointer;
  typedef ImageRegionConstIterator< PriorsImageType > PriorsImageIteratorType;

  /** Image Type and Pixel type for the images representing the membership of a
   *  pixel to a particular class. This image has arrays as pixels, the number of
   *  elements in the array is the same as the number of classes to be used. */
  typedef TInputVectorImage                               MembershipImageType;
  typedef typename MembershipImageType::PixelType         MembershipPixelType;
  typedef typename MembershipImageType::Pointer           MembershipImagePointer;
  typedef ImageRegionConstIterator< MembershipImageType > MembershipImageIteratorType;

  /** Image Type and Pixel type for the images representing the Posterior
   * probability of a pixel belonging to a particular class. This image has
   * arrays as pixels, the number of elements in the array is the same as the
   * number of classes to be used. */
  typedef VectorImage< TPosteriorsPrecisionType,
                       itkGetStaticConstMacro(Dimension) >                  PosteriorsImageType;
  typedef typename PosteriorsImageType::PixelType    PosteriorsPixelType;
  typedef typename PosteriorsImageType::Pointer      PosteriorsImagePointer;
  typedef ImageRegionIterator< PosteriorsImageType > PosteriorsImageIteratorType;

  /** Decision rule to use for defining the label. */
  typedef Statistics::MaximumDecisionRule DecisionRuleType;
  typedef DecisionRuleType::Pointer       DecisionRulePointer;

  typedef typename Superclass::DataObjectPointer DataObjectPointer;

  /** An image from a single component of the Posterior. */
  typedef itk::Image< TPosteriorsPrecisionType,
                      itkGetStaticConstMacro(Dimension) >                ExtractedComponentImageType;

  /** Optional Smoothing filter that will be applied to the Posteriors. */
  typedef ImageToImageFilter<
    ExtractedComponentImageType,
    ExtractedComponentImageType  >         SmoothingFilterType;

  typedef typename SmoothingFilterType::Pointer SmoothingFilterPointer;

  /** Set/Get the smoothing filter that may optionally be applied to the
   *  posterior image. */
  void SetSmoothingFilter(SmoothingFilterType *);
  itkGetConstMacro(SmoothingFilter, SmoothingFilterPointer);

  /** Set the priors image. */
  virtual void SetPriors(const PriorsImageType *);

  /** Number of iterations to apply the smoothing filter. */
  itkSetMacro(NumberOfSmoothingIterations, unsigned int);
  itkGetConstMacro(NumberOfSmoothingIterations, unsigned int);

  /** This is overloaded to create the Posteriors output image. */
  typedef ProcessObject::DataObjectPointerArraySizeType DataObjectPointerArraySizeType;
  using Superclass::MakeOutput;
  virtual DataObjectPointer MakeOutput(DataObjectPointerArraySizeType idx) ITK_OVERRIDE;

#ifdef ITK_USE_CONCEPT_CHECKING
  // Begin concept checking
  itkConceptMacro( UnsignedIntConvertibleToLabelsCheck,
                   ( Concept::Convertible< unsigned int, TLabelsType > ) );
  itkConceptMacro( PosteriorsAdditiveOperatorsCheck,
                   ( Concept::AdditiveOperators< TPosteriorsPrecisionType > ) );
  itkConceptMacro( IntConvertibleToPosteriorsCheck,
                   ( Concept::Convertible< int, TPosteriorsPrecisionType > ) );
  itkConceptMacro( InputHasNumericTraitsCheck,
                   ( Concept::HasNumericTraits< typename InputPixelType::ValueType > ) );
  itkConceptMacro( PosteriorsHasNumericTraitsCheck,
                   ( Concept::HasNumericTraits< TPosteriorsPrecisionType > ) );
  itkConceptMacro( PriorsHasNumericTraitsCheck,
                   ( Concept::HasNumericTraits< TPriorsPrecisionType > ) );
  itkConceptMacro( InputPriorsPosteriorsMultiplyOperatorCheck,
                   ( Concept::MultiplyOperator< typename InputPixelType::ValueType,
                                                PriorsPixelType, PosteriorsPixelType > ) );
  // End concept checking
#endif

protected:

  BayesianClassifierImageFilter();
  virtual ~BayesianClassifierImageFilter() ITK_OVERRIDE {}
  void PrintSelf(std::ostream & os, Indent indent) const ITK_OVERRIDE;

  virtual void GenerateData() ITK_OVERRIDE;

  virtual void GenerateOutputInformation(void) ITK_OVERRIDE;

  /** Compute the posteriors using the Bayes rule. If no priors are available,
   *  then the posteriors are just a copy of the memberships.
   *  Computes the labeled map for all combinations of conditions. */
  virtual void ComputeBayesRule();

  /** Normalize the posteriors and smooth them using a user-provided. */
  virtual void NormalizeAndSmoothPosteriors();

  /** Compute the labeled map based on the Maximum rule applied to the posteriors. */
  virtual void ClassifyBasedOnPosteriors();

  /** Get the Posteriors Image. */
  PosteriorsImageType * GetPosteriorImage();

private:

  ITK_DISALLOW_COPY_AND_ASSIGN(BayesianClassifierImageFilter);


  bool m_UserProvidedPriors;

  bool m_UserProvidedSmoothingFilter;

  SmoothingFilterPointer m_SmoothingFilter;

  unsigned int m_NumberOfSmoothingIterations;
};
} // end namespace itk

#ifndef ITK_MANUAL_INSTANTIATION
#include "itkBayesianClassifierImageFilter.hxx"
#endif

#endif